The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.
In a video data compression apparatus, intra encoding is performed by efficiently compressing residual signals, which are obtained by an efficient removal of a spatial redundancy in a frame or a block, through a transform technique, a quantization and an entropy coding process.
A typical intra encoding method includes an intra encoding technique through directionality-based extrapolation prediction of H.264/AVC.
The intra encoding through the directionality-based prediction of the H.264/AVC performs an extrapolation prediction by using left and upper pixels (A to M), which are previously encoded and decoded based on nine directions including a direct current or DC prediction (mode 2) as shown in FIG. 1. In other words, the extrapolation prediction can be performed in a manner that one pixel value in the block is filled by left and upper pixel values taken in a prediction direction. For example, in case of mode 0, since the extrapolation prediction is performed in a vertical direction, a 4×4 block of FIG. 1 has values filled with values A, B, C and D of the upper pixels in an arrow direction.
In recent times, in order to support ultra-high resolution, intra prediction methods have been proposed to use various directions and various block sizes. The inventor(s) has, however, noted that in these methods, in order to reflect a more precise texture direction in the prediction, a larger number of prediction modes should be calculated, and in particular, in order to process a large video having a high resolution, lengthy time is needed. The inventor(s) has noted that in addition, in these methods, as encoding and decoding results of the current block are used for the following block encoding and decoding, parallel processes of the encoding and decoding become difficult. That is, the inventor(s) has experienced that in order to decode the current block, reconstructed pixels of an upper block and a left block of the current block are needed, which holds encoding pixels of the current block until completion of the reconstruction of the previous block.
A technical solution has been proposed to provide a sub block level parallelism in a macro block. However, the inventor(s) has experienced that since the parallelism at best has a subsidiary block shape within the framework of the macro block, the parallelism cannot be easily implemented under current multidivisional structures. The inventor(s) has, further, experienced that as pixel values far from the current subsidiary block are used as extrapolation prediction values, compression efficiency will be reduced. The inventor(s) has noted that this brings a limiting factor to the video encoding and decoding through a parallel processing system of the state-of-the-art multi-core processor and graphic processing unit (GPU).